merge data script for FUSHAN

merge.data.FUSHAN.R

+### MERGE Fushan DATA
+### Edited by FH
+rm(list = ls()); source("./R/format.function.R"); library(reshape)
+
+#########################
+## READ DATA
+####################
+### read individuals tree data
+data.fushan <- load("./data/raw/DataFushan/fushan.RData"); data.fushan <- data.frame(fushan); rm(fushan)
+data.trait <- read.table("./data/raw/DataFushan/fs_trait_Kunstler.txt",header=T, sep = "\t") 
+colnames(data.trait) <- c("sp","sla","wd","seedmassmg","meanN","maxheightm")
+
+### read species names
+species.clean <- read.csv("./data/raw/DataFushan/Splist_Fushan_En.csv",stringsAsFactors=FALSE)
+
+######################################
+## MASSAGE TRAIT DATA
+############################
+## Obtain maximum height per species from data.trait; no sd available as we have only one observation for species
+data.max.height <- data.frame(sp = data.trait$sp, Max.height = log10(data.trait$maxheightm))
+data.fushan <- merge(data.fushan, data.max.height, by = "sp")
+
+##########################################
+## FORMAT INDIVIDUAL TREE DATA
+#############
+
+## change unit and names of variables to be the same in all data for the tree - DON'T HAVE A YEAR MEASUREMENT 
+data.fushan$G <- 10*(data.fushan$dbh2 -data.fushan$dbh1)/data.fushan$Interval ## diameter growth in mm per year
+data.fushan$year <- data.fushan$Interval ## number of year between measurement - MISSING
+data.fushan$D <- data.fushan[["dbh1"]]; ## diameter in cm
+data.fushan$dead <- as.numeric(data.fushan$status2 == "alive") ## dummy variable for dead tree 0 alive 1 dead
+data.fushan$plot <- (data.fushan[["PLOT_ID"]]) ## plot code - MISSING
+data.fushan$htot <- rep(NA,length(data.fushan[["dbh1"]]))## height of tree in m - MISSING
+data.fushan$tree.id <- data.fushan$tag ## tree unique id - MISSING
+data.fushan$sp.name <- NA; v <- species.clean$SP
+for(i in 1:length(unique(data.fushan$sp))) { 
+     sel.spp <- which(data.fushan$sp == unique(data.fushan$sp)[i])
+     data.fushan$sp.name[sel.spp] <- paste(species.clean$family[sel.spp],species.clean$genus[sel.spp], species.clean$epithet[sel.spp], sep = ".") }
+
+##########################################
+## CHANGE COORDINATE SYSTEM
+#############
+## DON'T KNOW!
+
+######################
+## ECOREGION
+###################
+## DON'T SEE DATA FOR THIS AT THE MOMENT, ALTHOUGH IT'S PROBABLY ALL IN ONE ECOREGION ANYWAY?
+
+######################
+## PERCENT DEAD
+###################
+## CANNOT DO THIS WITHOUT A PLOTID
+## compute numer of dead per plot to remove plot with disturbance
+perc.dead <- tapply(data.fushan[["dead"]],INDEX=data.fushan[["plot"]],FUN=function.perc.dead)
+# ## VARIABLE TO SELECT PLOT WITH NOT BIG DISTURBANCE KEEP OFTHER VARIABLES IF AVAILABLE (disturbance record)
+data.fushan <- merge(data.fushan,data.frame(plot=names(perc.dead),perc.dead=perc.dead), by = "plot", sort=FALSE)
+
+###########################################################
+### PLOT SELECTION FOR THE ANALYSIS
+###################
+## Remove data with dead == 1
+table(data.fushan$dead)
+data.fushan <- data.fushan[data.fushan$dead == 1,]
+
+colnames(data.fushan)[c(3,1,11,13)] <- c("sp","plot","w")
+vec.abio.var.names <-  c("MAT","MAP")
+vec.basic.var <-  c("tree.id","sp","sp.name","plot","D","G","dead","year","htot","gx","gy","perc.dead")
+data.tree <- subset(data.fushan,select=c(vec.basic.var,vec.abio.var.names))
+
+##############################################
+## COMPUTE MATRIX OF COMPETITION INDEX WITH SUM OF BA PER SPECIES IN EACH PLOT in m^2/ha without the target species
+###########################
+## CANNOT DO WITHOUT A PLOT ID
+data.BA.SP <- BA.SP.FUN(id.tree=as.vector(data.fushan[["tree.id"]]), diam=as.vector(data.fushan[["D"]]),
+	sp=as.vector(data.fushan[["sp"]]), id.plot=as.vector(data.fushan[["plot"]]),
+	weights=1/(pi*(0.5*data.fushan[["dbh1"]])^2), weight.full.plot=NA)
+
+## change NA and <0 data for 0
+data.BA.SP[is.na(data.BA.SP)] <- 0; data.BA.SP[,-1][data.BA.SP[,-1]<0] <- 0
+
+### CHECK IF sp and sp name for column are the same
+if(sum(!(names(data.BA.SP)[-1] %in% unique(data.fushan[["sp"]]))) >0) stop("competition index sp name not the same as in data.tree")
+
+#### compute BA tot for all competitors
+BATOT.COMPET <- apply(data.BA.SP[,-1],1,sum,na.rm=TRUE)
+data.BA.SP$BATOT.COMPET <- BATOT.COMPET; rm(BATOT.COMPET)
+### create data frame
+names(data.BA.SP) <- c("tree.id",names(data.BA.SP)[-1])
+data.BA.sp <- merge(data.frame(tree.id=data.fushan[["tree.id"]],ecocode=data.fushan[["ecocode"]]),data.BA.SP,by="tree.id",sort=FALSE)
+## test
+if(sum(!data.BA.sp[["tree.id"]] == data.tree[["tree.id"]]) >0) stop("competition index not in the same order than data.tree")
+
+## save everything as a list
+list.canada <- list(data.tree=data.tree,data.BA.SP=data.BA.sp,data.traits=data.traits)
+save(list.spain,file="./data/process/list.canada.Rdata")
+